Clinical Biochemistry 50 (2017) 174–180

Contents lists available at ScienceDirect

Clinical Biochemistry

journal homepage: www.elsevier.com/locate/clinbiochem

Analytical evaluation of a new point of care system for measuring cardiac

Troponin I
Danielle WM Kemper a,⁎, Veronique Semjonow a, Femke de Theije a, Diederick Keizer a, Lian van Lippen a,
Johannes Mair b, Bernadette Wille b, Michael Christ c, Felicitas Geier c, Pierre Hausfater d, David Pariente d,
Volkher Scharnhorst e, Joyce Curvers e, Jeroen Nieuwenhuis a
a
Philips Handheld Diagnostics, Eindhoven, The Netherlands
b
Department of Internal Medicine III — Cardiology and Angiology, Medical University of Innsbruck, Innsbruck, Austria
c
Department of Emergency and Critical Care Medicine, Paracelsus Medical University, Nuernberg General Hospital, Nuernberg, Germany
d
Emergency Department, Hôpital Pitié-Salpêtrière, AP-HP et Sorbonne Universités UPMC Univ-Paris 06, France
e
Clinical laboratory, Catharina Ziekenhuis Eindhoven and Technical University Eindhoven, Eindhoven, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: Point-of-care cardiac troponin testing with adequate analytical performances has the potential to
Received 20 September 2016 improve chest pain patients flow in the emergency department. We present the analytical evaluation of the
Received in revised form 31 October 2016 newly developed Philips Minicare cTnI point-of-care immunoassay.
Accepted 9 November 2016 Design & methods: Li-heparin whole blood and plasma were used to perform analytical studies. The sample
Available online 12 November 2016
type comparison study was performed at 4 different hospitals. The 99th percentile upper reference limit (URL)
study was performed using Li-heparin plasma, Li-heparin whole blood and capillary blood samples from 750
Keywords:
Emergency medicine
healthy adults, aging from 18 to 86 years.
Chest pain Results: Limit of the blank, limit of detection and limit of quantitation at 20% coefficient of variation (CV)
Acute myocardial infarction were determined to be 8.5 ng/L, 18 ng/L and 38 ng/L respectively without significant differences between
Cardiac Troponin-I whole blood and plasma for LoQ. Cross-reactivity and interferences were minimal and no high-dose hook was
Point-of-care observed. Total CV was found to be from 7.3% to 12% for cTnI concentrations between 109.6 and 6135.4 ng/L.
CV at the 99th percentile URL was 18.6%. The sample type comparison study between capillary blood, Li-
heparin whole blood and Li-heparin plasma samples demonstrated correlation coefficients between 0.99 and
1.00 with slopes between 1.03 and 1.08. The method comparison between Minicare cTnI and Beckman Coulter
Access, AccuTnI + 3 demonstrated a correlation coefficient of 0.973 with a slope of 1.09. The 99th percentile
URL of a healthy population was calculated to be 43 ng/L with no significant difference between genders or sam-
ple types.
Conclusions: The Minicare cTnI assay is a sensitive and precise, clinical usable test for determination of cTnI
concentration that can be used in a near-patient setting as an aid in the diagnosis of acute myocardial infarction.
© 2016 The Authors. Published by Elsevier Inc. on behalf of The Canadian Society of Clinical Chemists. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Measurement of cardiac Troponin-I or Troponin-T (cTnI, cTnT) con-

Abbreviations: AA, amino acid; AMI, acute myocardial infarction; CI, confidence
interval; CLSI, clinical laboratory standards institute; COPD, chronic obstructive
pulmonary disease; cTn, cardiac Troponin; cTnC, cardiac troponin C; cTnI, cardiac
Troponin I; CV, coefficient of variation; ED, emergency department; eGFR, estimated
glomerular filtration rate; HAMA, human anti-mouse antibody; Li-heparin, lithium hepa-
rin; LoB, limit of the blank; LoD, limit of detection; LoQ, limit of quantitation; NSTE-AMI,
non ST-segment elevation acute myocardial infarction; NT pro BNP, N-terminal pro
brain natriuretic peptide; POC, point-of-care; RF, rheumatoid factor; RFID, radio-
frequency identification; TAT, turnaround time; URL, upper reference limit.
⁎ Corresponding author at: Philips Handheld Diagnostics, High Tech Campus 29,
5656AE Eindhoven, The Netherlands.
E-mail address: Danielle.Kemper@philips.com (D.W.M. Kemper).
centration in blood is required for assessment of patients suspected of
Non ST-segment elevation acute myocardial infarction (NSTE-AMI) to
support or exclude a diagnosis of acute myocardial infarction (AMI)
[1,2,3]. Measuring cTnI at the point of care (next to the patient) with a
short turnaround time (TAT) has the potential to improve patients
flow in the emergency department (ED), enabling rapid clinical decision
making. A patient blood sample can be withdrawn and directly tested
by a doctor, a nurse or a paramedic to provide cTnI concentration during
clinical examination, rather than having to wait at least 1 h for laborato-
ry results [4,5]. A point-of-care (POC) cTn assay may allow the institu-
tion to meet the 1-h TAT criteria for the measurement of cTn from the

http://dx.doi.org/10.1016/j.clinbiochem.2016.11.011
0009-9120/© 2016 The Authors. Published by Elsevier Inc. on behalf of The Canadian Society of Clinical Chemists. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).
 D.W.M. Kemper et al. / Clinical Biochemistry 50 (2017) 174–180
guidelines [2,6]. Ideally, the analytical performance of a POC device for
cTn testing should not differ from that provided by the central laborato-
ry system [6]. We here present the results of the analytical evaluation of
the novel Minicare cTnI POC assay from Philips Electronics.
2. Materials and methods
2.1. Instrumentation
The Philips Minicare cTnI consists of a handheld instrument and
plastic disposable cartridge. The system makes use of the Philips
Magnotech technology, which is based on the precisely controlled mo-
tion of magnetic particles (beads) in a small sample volume (typically
30 μL). The same magnetic particles also serve as labels that are detected
using frustrated total internal reflection (FTIR) imaging [7,8].
2.2. Design of the minicare cTnI assay
The Minicare cTnI assay is a homogeneous sandwich immunoassay.
The traditional liquid manipulation steps of an immunoassay have been
replaced by magnetically controlled movements of magnetic nanoparti-
cles within a stationary liquid (see Fig. 1). The magnetic beads carry an-
tibodies directed against cTnI whereas the other side of the sandwich is
formed by antibodies printed on the bottom of the cartridge (the sensor
surface).
A droplet of sample (30 μL) of whole blood or plasma is applied to
the cartridge and the reaction chamber fills by capillary action. Red
blood cells are retained by an integrated separation membrane to pre-
vent impact on the assay. About 2 μL of plasma will be extracted and
the microfluidic design ensures that precisely 0.25 μL is metered into
the reaction chamber. In the first phase of the assay, beads coated
with antibody capture cTnI molecules in the sample. Subsequently,
magnetic fields gradients are engaged to transport the particles rapidly
to the sensor surface towards immobilized antibodies able to capture
the troponin-bearing nanobeads. Thereafter, a sequence of finely
tuned magnetic pulses is applied to facilitate optimal binding and
mixing of the beads containing cTnI molecules at the antibody-
functionalized surface. After the beads reacted with the sensor surface,
un-bound and non-specifically bound beads are rapidly removed with
a magnetic wash by applying a magnetic field gradient oriented away
from the detection surface [7,8].
2.3. Selection of antibodies
Antibodies are applied to the magnetic beads and onto the plastic
surface of the cartridge to form both parts of the sandwich, as described
above. The choice of antibodies is crucial for assay performance. Mouse-
monoclonal antibodies were selected for the Philips Minicare cTnI assay.
The primary anti-cTnI mouse-monoclonal antibody is directed against
Fig. 1. Depiction of the reaction chamber and actuation magnets showing the assay processes: analyte binding by beads bearing anti-cTnI antibodies (top and bottom magnets off), bead
binding to the sensor surface (bottom magnet on) and magnetic removal of free and weakly bound beads (top magnet on).
175
the stable region of the cTnI molecule (amino acid (AA) 41–49) and
has been covalently bound to the magnetic beads. A mixture of three
secondary antibodies have been attached to the sensor surface by
physisorption, which consists of two anti-cTnI antibodies with epitopes
in the range AA 20–100 and a single anti-cTnC antibody, to optimize
measurement of total cTnI including cTnI-TnC complexes.
2.4. Standardization
Due to the lack of a suitable commutable primary calibrator for cTnI
immunoassay standardization, new cTnI standards have been devel-
oped from pooled native human samples. These standards are dose-
assigned on the Beckman Coulter Access 2, AccuTnI+3 as a reference
method. Primary calibrators are prepared, based on the standard refer-
ence material of the National Institute of Standards and Technology
(NIST SRM) 2921 Human Cardiac Troponin Complex, and are dose-
assigned on the native standards. Additionally, secondary calibrators
are prepared from NIST SRM 2921 and dose assigned using the primary
calibrators. Calibration parameters were obtained from dose-response
curves and programmed in the radio-frequency identification (RFID)
of each cartridge.
2.5. Study samples
For the precision, detection capability, linearity, high-dose hook ef-
fect, cross-reactivity and interferences studies, left-over Li-Heparin
blood samples from patients from the Canisius-Wilhelmina hospital
(CWZ) in Nijmegen, The Netherlands, and Li-Heparin blood samples
from healthy volunteers (informed consent obtained) from Sanquin
Blood bank in Nijmegen, the Netherlands were obtained. From these
samples negative and high cTnI Li-heparin plasma pools were selected
to be able to prepare sample pools with different levels of cTnI as spec-
ified for each study.
For the method comparison and sample type comparison studies,
three sample types (capillary whole blood from finger stick, Li-heparin
whole blood and Li-heparin plasma) were collected for each patient at
four European hospitals (Medizinische Universitaet Innsbruck, Austria,
Klinikum Nurnberg Germany, Catharina Ziekenhuis Eindhoven, The
Netherlands and Hôpital de la Pitié-Salpêtrière Paris, France) participat-
ing to the Lab2Go project. Lab2Go is a European Union funded multicen-
ter Research and Development project involving several hospitals in the
European Union. Patients were selected to represent the range of cTnI
concentrations likely to be encountered in clinical practice, covering
the measurement range of the Minicare cTnI. Samples were analysed
on Minicare cTnI by Minicare-trained users (nurses, research assistants)
within 2 h after blood drawn. Li-heparin plasma samples were centri-
fuged a second time and transferred to a new container before freezing
at b−55 °C. Frozen samples were sent to the core lab at Philips on dry
ice for parallel testing on the Beckman Coulter AccuTnI + 3 assay and
176 D.W.M. Kemper et al. / Clinical Biochemistry 50 (2017) 174–180
the Minicare cTnI for the method comparison study. This study has been
approved by all local ethical committees and informed consent was ob-
tained from patients prior to their participation in the study.
For the 99th percentile upper reference limit (URL) study Li-heparin
whole blood, capillary whole blood and Li-heparin plasma samples
were obtained from healthy volunteers at PRA Health Sciences
Zuidlaren, The Netherlands. Informed consent was obtained from all
volunteers prior to their participation in the study.
2.6. Precision
To assess the Minicare cTnI imprecision according to CLSI EP05-A3
recommendations [9] three Li-heparin plasma pools with different
levels of cTnI distributed over the measuring range were tested in two
runs per day on twenty different days, using two different lots of
Minicare cTnI cartridges.
2.7. Detection capability
Limit of the blank (LoB) and limit of detection (LoD) in Li-heparin
plasma and limit of quantitation (LoQ) in Li-heparin whole blood and
Li-heparin plasma for the Minicare cTnI were established in accordance
with CLSI document EP17-A2 [10]. A total of four blank Li-heparin plas-
ma pools for LoB, four Li-heparin plasma pools with cTnI concentrations
of 15, 20, 25 and 30 ng/L for LoD and four Li-heparin plasma pools with a
cTnI concentration between 45 and 55 ng/L for LoQ were measured on
three different days, in five-fold per day, on two cartridge lots giving
120 measurements for LoB, LoD and LoQ in Li-heparin plasma. Further-
more, for measuring LoQ in whole blood, 16 Li-heparin whole blood
samples (concentration between 10 and 200 ng/L) were measured in
20-fold on ten analyzers in parallel, on one cartridge lot and on multiple
days to determine LoQ in Li-heparin whole blood.
2.8. Linearity on dilution
According to CLSI EP06-A recommendation [11], 11 Li-heparin plas-
ma pools with different levels of cTnI were prepared from a high cTnI Li-
heparin plasma pool and a negative Li-heparin plasma pool with steps
of 10% dilution. The pool with the highest cTnI level and the pool with
the lowest cTnI level were measured in five-fold. All the other pools
were measured in three-fold. All tests were performed on one Minicare
cTnI cartridge lot.
2.9. High-dose hook effect
11 dilutions blends prepared from the NIST standard of human cTnI
(SRM 2921) and a negative Li-heparin plasma pool were tested. All
high-dose hook samples were measured in duplicate on one Minicare
cTnI cartridge lot.
2.10. Cross-reactivity and interferences
Various blood components (hemoglobin, bilirubin, triglycerides, lec-
ithin, human anti-mouse antibodies (HAMA)) and drugs (allopurinol,
acetaminophen, ampicillin, ascorbic acid, acetylsalicylic acid, atenolol,
caffeine, captopril, digoxin, dopamine-HCL, erythromycin, furosemide,
methyldopa, niphedipine, phenytoin, theophylline, verapamil) were
tested for interference and human skeletal troponin I, human cardiac
troponin T, human cardiac troponin C and human skeletal troponin T
(Hytest 8T25, 8T13, 8T57 and 8T24 respectively) for cross-reactivity ac-
cording to CLSI EP07-A2 recommendations [12] in a Li-heparin plasma
pool with a cTnI concentration of 105 ng/L. As a reference, only the dilu-
tion matrix of the substances was tested in the Li-heparin plasma pool.
Results for samples enriched with these possible interferents that were
within 10% of the results for the control samples were considered ac-
ceptable. Four endogenous interfering substances (human albumin,
globulin, rheumatoid factor (RF) and total protein) were tested on inter-
ference in samples from a normal population (without spiking) on
Minicare cTnI and a comparative method (Beckman Coulter Access 2,
AccuTnI+3).
2.11. Method comparison
The Minicare cTnI was compared with the Beckman Coulter Access
2, AccuTnI+3 by testing and comparing 119 Li-Heparin plasma samples
in accordance with CLSI document EP09-A3 [13].
2.12. Sample type comparison
The sample type comparison study was conducted on 138 paired pa-
tient samples from 4 hospitals, all part of the Lab2Go consortium, in ac-
cordance with CLSI EP09-A3 recommendations [13]. From each patient
capillary whole blood from finger stick, Li-heparin whole blood and Li-
heparin plasma from venous puncture were assayed in duplicate on
Minicare cTnI.
2.13. 99th percentile upper reference limit (URL)
A single site study was performed at PRA Health Sciences,
Stationsweg 163, 9471 GP, Zuidlaren, The Netherlands. The health con-
dition of 848 apparently healthy adults was screened based on a ques-
tionnaire, on the measurement of a cardiac marker (NTproBNP on the
Siemens Immulite 2000) and a marker for kidney function (creatinine
to estimate Glomerular Filtration Rate).
Volunteers with a personal history of AMI or other cardiac vascular
diseases, COPD, immunological disease, diabetes, hypertension, renal
disease, drug-of-abuse or cancer in the last 5 years were excluded
from the study. The cutoff values used to exclude patients based on
NTproBNP test were taken from the package insert of the Siemens
Immulite 2000 NT-proBNP and were N125 pg/mL for subjects
b75 years old and N450 pg/mL for subjects 75 years old or older. Accep-
tance criteria for eGFR was above 60 mL/min/1.73 m2. In total 750
healthy volunteers (373 males and 377 females; age range from 18 to
86 years) were qualified as final study population for the 99thpercentile
URL study. Males and females were equally distributed over five age
groups (18–30, 31–40, 41–50, 51–60 and N60). The 99th percentile
URL for the Minicare cTnI was determined by testing capillary whole
blood, Li-Heparin whole blood and Li-Heparin plasma samples and
has been calculated per sample type and per gender.
2.14. Statistical analysis
The LoB was calculated using the nonparametric method with for-
mula: Rank position = 0.5 + B ∗ 0.95 (where B is the number of repli-
cates.). The LoD was calculated using the formula: LoD = LoB + cpSDL
where cp is a multiplier to give the 95th percentile of a normal distribu-
tion (equal to 1.653) and SDL represents the standard deviation (SD) of
all the results of the pooled samples. For LoQ a coefficient of variation
(CV) profile of 16 Li-heparin venous blood samples with concentrations
between 10 and 200 ng/L was calculated to determine the LoQ at a total
CV of 20%. The calculations were modelled using the equation CV = a
(concentration)b + c to model the CV's with concentration, where a
and b are constants to fit and, in addition, background CV c was
added. This nonlinear model was solved with Gauss-Newton. This
model was furthermore used to calculate the CV at 99th percentile
URL. The linearity was checked using IVDfit as statistical analysis tools.
A 2nd polynomial fit was calculated with a weight factor (1/SD^2) as
described in CLSI guideline EP06-A section 5.3.3 [11]. Method compari-
son regression analyses were performed using the Deming regression
method. The sample type comparison analysis was performed accord-
ing to the Passing and Bablock linear regression procedure [14]. The
agreement between the two sample types was assessed according to
 D.W.M. Kemper et al. / Clinical Biochemistry 50 (2017) 174–180
the method as described by Bland and Altman [15]. For these analyses
methods and additional basic statistics Analyze-it for Microsoft Excel
was used as software tool. Since no common statistical distribution de-
scribed the data set properly at percentiles over 98–99, the 99th percen-
tile was determined using the non-parametric method (PROC
UNIVARIATE with interpolation option (1) in SAS® statistical software,
i.e. linear interpolation between the data points closest to the 99th per-
centile) as described in CLSI standard EP28-A3c [16]. Furthermore, as
described in this guideline, 90% CI's were calculated based on binomial
probabilities. The lower and upper ranks corresponding to the confi-
dence limits are symmetric in ranks and chosen so that their values
are as close to the 99th percentile as possible while satisfying the cover-
age requirement.
3. Results
3.1. Precision
For the 3 Li-heparin plasma pools 80 replicates were measured. The
total imprecision of the Li-heparin plasma pools with concentrations
between 109.6 and 6135.4 ng/L and across two lots was found to be be-
tween 7.3% and 12.0%. At low cTnI level the total imprecision for lot 1
was 11.6% and for lot 2 it was 12.0%. In the medium range of cTnI
level, total imprecision for lot 1 was 9.6% and for lot 2 was 8.4%. For
the high cTnI level sample, the total imprecision for lot 1 was 7.3% and
lot 2 it was 7.7%.
3.2. Detection capability
For the four samples for LoB, a total of 60 replicates were measured
and ranked on increasing concentration. The LoB for lot 1 and lot 2 was
found to be 6.5 ng/L and 8.5 ng/L respectively. The highest value of LoB
was used to calculate the LoD. The LoD was determined to be 18 ng/L
and 17 ng/L for lot 1 and lot 2 respectively. For Li-heparin whole blood
a CV profile was calculated in order to determine the LoQ at 20%CV
(Fig. 2). The LoQ for Li-heparin whole blood was found to be 38 ng/L
[95% CI 28.3–47.7 ng/L]. The LoQ at 20%CV for Li-heparin plasma was de-
termined to be 37 and 29 ng/L for lot 1 and lot 2 respectively.
3.3. Linearity
The Minicare cTnI demonstrated linearity with a maximum devia-
tion between a linear and non-linear fit of ≤15% throughout the mea-
sured range up to and including 8126 ng/L.
CV= a (concentration)b+c
a = 434.3599707
b = -0.975450351
c = 7,5
Fig. 2. CV profile for Li-heparin whole blood. On the horizontal axis the Average Minicare
cTnI concentration in ng/L is depicted. On the vertical-axis the CV in (%) is depicted.
177
3.4. High-dose Hook effect
No High-dose hook effect was found for samples up to and including
a cTnI concentration of 2,000,000 ng/L.
3.5. Cross-reactivity and interference
None of the tested blood components and/or drugs showed interfer-
ence beyond 90% and 110%. The interference of four endogenous inter-
fering substances albumin, globulin, total protein and RF were
determined in natural plasma samples via a method comparison
study, showing that there is no predictive relationship between the in-
terfering factor content and specific bias of the cTnI results up to a con-
centration of 4.95 g/dL, 1.75 g/dL, 8.09 g/dL and 45.3 IU/mL respectively.
The observed percentage of cross-reactivity of the Minicare cTnI to
human skeletal troponin I, human cardiac troponin T, human cardiac
troponin C and human skeletal troponin T was 0.00045%, 0.00664%,
0.00256% and 0.00016% respectively.
3.6. Method comparison
In Fig. 3, the Deming regression is shown for the comparison be-
tween Beckman Coulter Access, AccuTnI + 3 and Minicare cTnI per-
formed on Li-heparin plasma samples from 119 patients. The Pearson
correlation was good: over all measurements the correlation coefficient
was 0.973 [95% CI 0.961–0.981]. The slope was 1.09 [95% CI 0.97–1.20]
with an intercept at 75.18 ng/L [95% CI 33.7–116.7]. No outlier was
discarded from the data set.
3.7. Sample type comparison
Sample types were prepared from blood samples from 122 patients
with cTnI concentrations that spanned the full measurement range of
the Minicare cTnI. In Fig. 4 the Passing-Bablock fit and Bland-Altman fig-
ures are shown for the comparison between the three sample types. The
sample type correlations are reported in Table 1. Samples with a con-
centration below LoD (18 ng/L) were excluded from the data set in
the Passing-Bablock and correlation plot and samples with a concentra-
tion below LoQ whole blood (38 ng/L) were excluded in the Bland-
Altman plot. The correlation between the sample types was excellent.
Over all measurements, with cTnI values covering 18–7000 ng/L, the
sample type comparison for Li-heparin venous whole blood versus cap-
illary whole blood, Li-heparin venous whole blood versus Li-heparin
Fig. 3. Method comparison on 119 Li-heparin plasma samples using Analyse-it for analysis.
178 D.W.M. Kemper et al. / Clinical Biochemistry 50 (2017) 174–180

Fig. 4. Correlation between cTnI concentrations as obtained with the Minicare cTnI analyzer in Li-heparin venous whole blood and Li-heparin venous plasma samples (A), Li-heparin
venous plasma and capillary whole blood samples (C) and Li-heparin venous whole blood and capillary whole blood samples (E) (122 samples for each pair). Bland-Altman graphs for
venous whole blood vs plasma samples (B), plasma vs capillary whole blood samples (D) and venous whole blood vs capillary whole blood samples (F) (104 samples for each pair).

plasma and Li-heparin plasma versus capillary whole blood showed R capillary whole blood, Li-heparin venous whole blood versus Li-
values of 0.995, 0.996 and 0.990 and slopes of 1.08, 1.03 and 1.05 respec- heparin plasma and Li-heparin plasma versus capillary whole blood
tively. The bias measured for Li-heparin venous whole blood versus were 7.7%, 2.2% and 5.5%, respectively.

Table 1
Correlation coefficients (R) and slopes with 95% CI for measurements of cTnI in different sample types using the Minicare I-20 analyzer. Data based on 122 patient samples in a range of 18–
7000 ng/L and Bland-Altman data (mean difference with 95% CI together with lower and upper limits of agreement (LoA)) using Analyse-it for analysis with 104 samples.

Sample types R Slope (95% CI) Mean difference 95% CI lower LoA upper LoA

Li-heparin whole blood vs. Li-heparin plasma 1.00 1.03 (1.01–1.05) 2.2% −0.2%–4.6% −22.0 26.5
Li-heparin plasma vs. capillary whole blood 0.99 1.05 (1.02–1.08) 5.5% 2.9%–8.0% −20.7 31.6
Li-heparin whole blood vs. capillary whole blood 1.00 1.08 (1.07–1.10) 7.7% 5.5%–9.9% −14.5 29.9
 D.W.M. Kemper et al. / Clinical Biochemistry 50 (2017) 174–180
Fig. 5. Distribution of Minicare cTnI values of capillary blood from healthy individuals resulting from the 99th percentile URL study.
3.8. 99th percentile URL
In Fig. 5 the distribution of the Minicare cTnI levels of the healthy in-
dividuals resulting from the 99th percentile URL study is shown. Mea-
surements on capillary blood are shown as an example. Other sample
types show similar distributions. The presence of outliers was verified
carefully [17]. As the dataset was clearly non-normal and transforma-
tion to a normal distribution was not possible using standard transfor-
mations, Tukey's rule [16] was not applicable. Application of Dixon's
test [16] revealed that the highest data point for all data sets was a po-
tential outlier. However, since for this particular subject all three sepa-
rate measurements were higher and the subject was qualified as a
healthy person, i.e. belonged to the healthy population, these data
points were not discarded.
The 99th percentile URL of Minicare cTnI for 750 healthy volunteers
(373 males and 377 females, age range from 18 to 86 years) was calcu-
lated at 59 ng/L using capillary blood, 39 ng/L using Li-heparin whole
blood and 41 ng/L using Li-heparin plasma (see Table 2).No significant
influence of one of the three tested samples types on the overall 99th
percentile cut-off values was observed so that the same 99th percentile
URL could be used for the three sample types. The combined 99th per-
centile URL value for all samples types was 43 ng/L (90% CI: 35–
61 ng/L). There were also no significant differences between cTnI values
of male and female subjects for any of the sample types, allowing com-
bining male and female cTnI values for the determination of the
Minicare cTnI 99th percentile URL. Testing for equality of the 99th per-
centile URL's of all six groups (three blood types combined with two
genders) was done by using Pearson's chi-square [18] and showed no
significant difference (Pearson chi2(5) = 4.8570; Pr = 0.434). The
99th percentile URL for all sample types and genders was established
at 43 ng/L. From the CV profile of Li-heparin whole blood (Fig. 2), the
CV at 99th percentile URL was calculated to be 18.6%.
Table 2
Summary 99th percentile URL values Minicare cTnI (ng/L).
Sample type
Capillary whole blood
Li-heparin venous whole blood
Li-heparin plasma
Capillary whole blood and Li-heparin venous whole blood combined
Capillary whole blood and Li-heparin plasma combined
Li-heparin venous whole blood and Li-heparin plasma combined
Overall
179
4. Discussion
The study results show that the Minicare cTnI is a clinically usable
cTnI POC test which can accurately and rapidly measure cTnI near the
patient with a turn-around-time of b10 min, which is somewhat faster
than currently available POC cTnI assays which usually need 10–
20 min test time [19], making the Minicare cTnI one of the fastest POC
devices for cTn testing. Unlike other cTn assays which need at least
90 μL of sample, the Minicare cTnI only requires a single droplet of
blood and is the first device on the market that has been evaluated care-
fully for capillary sample testing. This allows for minimally invasive
blood collection via finger prick, giving the assay a good flexibility in
use. The Minicare cTnI needs no sample preparation on forehand,
which is needed for some other cTnI POC assays, thereby reducing TAT
and the possibility of user errors [19]. Results will be available to the
treating physician at bedside together with the ECG after history taking
which potentially accelerates the ACS patient's pathway and reduces ED
crowding [20,21]. Further reduction in TAT could be achieved by
performing the first measurement in the ambulance; the ruggedness
of the device seems suitable for this use environment and studies have
been initiated to investigate this workflow.
For LoQ at 20%CV no significant difference between Li-heparin whole
blood and Li-heparin plasma was observed. The sample type comparison
study between capillary whole blood, Li-heparin whole blood and Li-
heparin plasma samples demonstrated an excellent correlation and indi-
cates that the three sample types are substantially equivalent. Thus,
sample types can be used interchangeable which is especially beneficial
in case of serial testing. Capillary sampling could also ease cTnI testing in
other medical fields, like the general practitioner's office, ambulance cars
and at the triage nurse in the ED. Method comparison between Minicare
cTnI and the reference cTnI assay (Beckman Coulter Access, AccuTnI+3)
demonstrated a very close correlation. Cross-reactivity and interferences
Overall (n = 750) Male (n = 373) Female (n = 377)
59 (90% CI: 40–88) 67 (90% CI: 30–387) 52 (90% CI: 30–88)
39 (90% CI: 25–61) 51 (90% CI: 21–329) 33 (90% CI: 23–59)
41 (90% CI: 28–93) 68 (90% CI: 23–350) 37 (90% CI: 24–93)
44 (90% CI: 33–63) 58 (90% CI: 35–165) 40 (90% CI: 25–59)
49 (90% CI: 36–69) 65 (90% CI: 37–165) 40 (90% CI: 28–64)
39 (90% CI: 28–59) 56 (90% CI: 30–101) 33 (90% CI: 24–43)
43 (90% CI: 35–61) 60 (90% CI: 37–72) 39 (90% CI: 30–49)
180 D.W.M. Kemper et al. / Clinical Biochemistry 50 (2017) 174–180
were minimal and no high-dose hook effect was observed. Since the
Minicare uses 2 solid phases (with one antibody on the magnetic bead
and the complimentary antibodies on the sensor surface) accessibility
to the target epitopes could be more challenging compared to systems
that only use one solid phase and have the other antibody free in solu-
tion. One can hypothesize that the potential negative steric impact from
using two solid phases can be partially mitigated by including additional
epitopes. This has been the rationale to evaluate configurations with an
antibody against cTnC in the spot. On average the signal increased
about 25–30% (so 25–30% more beads were bound to the spot for a com-
parable TnI concentration) when including the cTnC antibody to the spot.
It should be noted that the Minicare cTnI assay is still specific for cTnI (po-
tentially as part of the complex with cTnC) since the antibody on the bead
is directed against cTnI and hence a label can only be bound in the pres-
ence of cTnI. A potential downside of including a cTnC antibody in the
spot could be increased non-specific binding against cTnC for the assay.
However, in our studies cross-reactivity with cTnC was very low
(b0.005%), which might be explained by the good specificity of the anti-
body used on the bead. Based on the above it was decided to include a
cTnC antibody into our final assay configuration.
The CV at the 99th percentile URL of 43 ng/L, as measured conserva-
tively, on a truly healthy population, was calculated to be 18.6%. Accord-
ing to Jaffe et al. [22], LoQ below or equal to 20%CV at 99th percentile
URL leads to a clinically usable system for the diagnosis of AMI. The
99th percentile URL showed no significant difference between genders
or sample types in a large healthy population selected according to
the most recent recommendations for normal subjects to determine
the 99th percentile URL of a cTn assay [23]. These recommendations ad-
vise to include at least 300 male and 300 female subjects without clini-
cal history or known cardiovascular disease, no diabetes, age range
between 18 to above 70 years old, diversity in race/ethnicity, biomarker
negative for NT pro BNP and normal eGFR values.
In conclusion, the Minicare cTnI assay is a sensitive, fast and precise
test for determination of cTnI that can be used as an aid in the diagnosis
of AMI in a near-patient setting on capillary or Li-heparin venous whole
blood sample. This offers for the Minicare cTnI test the potential for
good clinical performance in the diagnosis of AMI, which is currently
evaluated in a large clinical study.
Acknowledgement
The European Union (project 621035) funded part of the studies by
supporting the following sites: Medizinische Universität Innsbruck,
Univ. Klinik für Innere Medizin III – Kardiologie und Angiologie (Inns-
bruck, Austria), Klinik für Notfall- und Internistische Intensivmedizin,
Klinikum Nürnberg Nord (Nurnberg, Germany), Hôpital Universitaire
La Pitié-Salpêtrière, Service des Urgences du Pr Riou (Paris, France)
and Catharina Ziekenhuis Eindhoven, Algemeen Klinisch Laboratorium
(Eindhoven, the Netherlands).
References
[1] K. Thygesen, J.S. Alpert, A.S. Jaffe, M.L. Simoons, B.R. Chaitman, H.D. White, The writ-
ing group on behalf of the joint ESC/ACCF/AHA/WHF task force for the universal def-
inition of myocardial infarction. Third universal definition of myocardial infarction,
Eur. Heart J. 33 (2012) 2551–2567.
[2] C.W. Hamm, J.P. Bassand, S. Agewall, J. Bax, E. Boersma, H. Bueno, et al., ESC Commit-
tee for practice guidelines. ESC guidelines for the management of acute coronary
syndromes in patients presenting without persistent ST-segment elevation: the
task force for the management of acute coronary syndromes (ACS) in patients pre-
senting without persistent ST-segment elevation of the European Society of Cardiol-
ogy (ESC), Eur. Heart J. 32 (23) (2011) 2999–3054.
[3] E.A. Amsterdam, N.K. Wenger, R.G. Brindis, D.E. Casey, T.G. Ganiats, D.R. Holmes,
et al., 2014 AHA/ACC guideline for the management of patients with non–ST-
elevation acute coronary syndromes: executive summary: a report of the
American College of Cardiology/American Heart Association Task Force on Practice
Guidelines, Circulation 130 (2014) 2354–2394.
[4] R. Bingisser, C. Cairns, M. Christ, P. Hausfater, B. Lindahl, J. Mair, et al., Cardiac tropo-
nin: a critical review of the case for point-of-care testing in the ED, Am. J. Emerg.
Med. 30 (8) (2012) 1639–1649.
[5] K.D. Rooney, U.M. Schilling, Point-of-care testing in the overcrowded emergency de-
partment — can it make a difference? Crit. Care 18 (2014) 692.
[6] D.A. Morrow, C.P. Cannon, R.L. Jesse, L.K. Newby, J. Ravkilde, A.B. Storrow, et al., Na-
tional academy of clinical biochemistry laboratory medicine practice guidelines:
clinical characteristics and utilization of biochemical markers in acute coronary syn-
dromes, Clin. Chem. 53 (2007) 552–574.
[7] D.M. Bruls, T.H. Evers, J.A. Kahlman, P.J. van Lankvelt, M. Ovsyanko, E.G. Pelssers,
et al., Rapid integrated biosensor for multiplexed immunoassays based on actuated
magnetic nanoparticles, Lab Chip 9 (24) (2009) 3504–3510.
[8] W.U. Dittmer, T.H. Evers, W.M. Hardeman, W. Huijnen, R. Kamps, P. de Kievit, et al.,
Rapid, high sensitivity, point-of-care test for cardiac troponin based on
optomagnetic biosensor, Clin. Chim. Acta 411 (2010) 868–873.
[9] Clinical Laboratory Standards Institute, Evaluation of Precision of Quantitative Mea-
surement Procedures; Approved Guideline—Third Edition. CLSI Document EP05-A3,
CLSI, Wayne (PA), 2014.
[10] Clinical Laboratory Standards Institute, Evaluation of Detection Capability for Clini-
cal Laboratory Measurement Procedures; Approved Guideline—Second Edition.
CLSI Document EP17-A2, Wayne (PA), CLSI, 2012.
[11] Clinical Laboratory Standards Institute, Evaluation of the Linearity of Quantitative
Measurement Procedure: A Statistical Approach; Approved Guideline. CLSI Docu-
ment EP06-A, Wayne (PA), CLSI, 2003.
[12] Clinical Laboratory Standards Institute, Interference Testing in Clinical Chemistry;
Approved Guideline—Second Edition. CLSI Document EP07-A2, Wayne (PA), CLSI,
2005.
[13] Clinical Laboratory Standards Institute, Measurement Procedure Comparison and
Bias Estimation Using Patient Samples; Approved Guideline—Third Edition. CLSI
Document EP09-A3, Wayne (PA), CLSI, 2013.
[14] H. Passing, W. Bablok, A new biometrical procedure for testing the equality of mea-
surements from two different analytical methods. Application of linear regression
procedures for method comparison studies in clinical chemistry, part I, J. Clin.
Chem. Clin. Biochem. 21 (1983) 709–720.
[15] J.M. Bland, D.G. Altman, Statistical methods for assessing agreement between two
methods of clinical measurement, Lancet. (1986) 307–310, i.
[16] Clinical Laboratory Standards Institute, Defining, Establishing, and Verifying Refer-
ence Intervals in the Clinical Laboratory; Approved Guideline—Third Edition. CLSI
Document EP28-A3c, Wayne (PA), CLSI, 2010.
[17] K.M. Eggers, F.S. Apple, L. Lind, B. Lindahl, The applied statistical approach highly in-
fluences the 99th percentile of cardiac troponin I, Clin. Biochem. 49 (2016)
1109–1112.
[18] W.D. Johnson, R.A. Beyl, J.H. Burton, C.M. Johnson, J.E. Romer, L. Zhang, Use of
Pearson's Chi-Square for testing equality of percentile profiles across multiple pop-
ulations, Open J of Stat 5 (2015) 412–420.
[19] E. Amundson, F.S. Apple, Cardiac troponin assays: a review of quantitative point-of-
care devices and their efficacy in the diagnosis of myocardial infarction, Clin. Chem.
Lab. Med. 53 (5) (2015) 665–675.
[20] C. Loten, J. Attia, C. Hullick, J. Marley, P. McElduff, Point of care troponin decreases
time in the emergency department for patient with possible acute coronary syn-
drome: a randomized controlled trial, Emerg. Med. J. 27 (2010) 194–198.
[21] P.S. Rahko, Rapid evaluation of chest pain in the emergency department, JAMA In-
tern. Med. 174 (1) (2014) 59–60.
[22] S. Jaffe, F.S. Apple, The third universal definition of myocardial infarction — moving
forward, Clin. Chem. 58 (12) (2012) 1727–1728.
[23] Y. Sandoval, F.S. Apple, The global need to define normality: the 99th percentile
value of cardiac troponin, Clin. Chem. 60 (3) (2014) 455–462.